Elasticsearch nodes in Elassandra multicluster

[banthit...@gmail.com]

Hi team,

I've a Elassandra Multi datacenter DC1(running 2 nodes) and DC2(running one node).

three of them are perfectly sharing Cassandra keyspaces but for Elasticsearch indices, I think I have to do a manual re-mapping from Cassandra schemas to Elasticsearch schemas due to the mismatched version. (5.5 and 6.x).

The structure

DC1

===============

x.x.x.x: version 5.5

x.x.x.x: version 6.2

DC2

===============

x.x.x.x: version 6.4

So I have a couple of questions,

1. Am I right about re-mapping schema? If that true, I have to list all the Cassandra schema and transform to Elasticsearch schema.

2. When joining the datacenter with multicluster, the node in DC2 unable to see the others. from  `curl -X GET "localhost:9200/_cluster/health?pretty"` on itself only shows 1 node discovered and no shards at all.

root@ns3053180:~# curl -X GET "localhost:9200/_cluster/health?pretty"

________________________________________________________________________________

{

  "cluster_name" : "Test Cluster@Elassandra",

  "status" : "green",

  "timed_out" : false,

  "number_of_nodes" : 1,

  "number_of_data_nodes" : 1,

  "active_primary_shards" : 0,

  "active_shards" : 0,

  "relocating_shards" : 0,

  "initializing_shards" : 0,

  "unassigned_shards" : 0,

  "delayed_unassigned_shards" : 0,

  "number_of_pending_tasks" : 0,

  "number_of_in_flight_fetch" : 0,

  "task_max_waiting_in_queue_millis" : 0,

  "active_shards_percent_as_number" : 100.0

}

________________________________________________________________________________

but for the other two (in different datacenter) also show only two of themselves (not found the node above)

 {

  "cluster_name" : "Test Cluster@Elassandra",

  "status" : "green",

  "timed_out" : false,

  "number_of_nodes" : 2,

  "number_of_data_nodes" : 2,

  "active_primary_shards" : 0,

  "active_shards" : 0,

  "relocating_shards" : 0,

  "initializing_shards" : 0,

  "unassigned_shards" : 0,

  "delayed_unassigned_shards" : 0,

  "number_of_pending_tasks" : 0,

  "number_of_in_flight_fetch" : 0,

  "task_max_waiting_in_queue_millis" : 0,

  "active_shards_percent_as_number" : 100.0

}

the question is, do I did something wrong? should all of the nodes seeing each other?

3. I've installed Kibana in node in DC2,  but the Kibana server haven't came up

It always say "Kibana server is not ready yet" . I did everything that I can to bring the server up and I've set the configuration in kibana.yml. Unfortunately, the server still down with only an error 

{"type":"log","@timestamp":"2020-08-24T07:49:23Z","tags":["fatal","root"],"pid":25623,"message":"{ Error: Request Timeout after 30000ms\n    at /usr/share/kibana/node_modules/elasticsearch/src/lib/transport.js:355:15\n    at Timeout.<anonymous> (/usr/share/kibana/node_modules/elasticsearch/src/lib/transport.js:384:7)\n    at ontimeout (timers.js:436:11)\n    at tryOnTimeout (timers.js:300:5)\n    at listOnTimeout (timers.js:263:5)\n    at Timer.processTimers (timers.js:223:10)\n  status: undefined,\n  displayName: 'RequestTimeout',\n  message: 'Request Timeout after 30000ms',\n  body: undefined }"}

The error doesn't tell much so I have no any idea what should I go next.

Thanks in advance for all the helps

Regards,

[Vincent Royer]

You must run the same major version in a datacenter, so DC1 cannot work like that.

You must use datacenter group (see doc) to run independent elasticsearch cluster in each DC, and recreate the mapping when you upgrade the major version.

 

--
You received this message because you are subscribed to the Google Groups "Elassandra User Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elassandra+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elassandra/362ba7be-138a-4657-b952-369a03591b47n%40googlegroups.com.